Mission,Motivation,and the Active Decision to Work for a Social Cause

Abstract

The mission of a job affects the type of worker attracted to an organization but may also provide incentives to an existing workforce. We conducted a natural field experiment with 246 short-term workers. We randomly allocated some of these workers to either a prosocial or a commercial job. Our data suggest that the mission of a job has a performance-enhancing motivational impact on particular individuals only, those with a prosocial attitude. However, the mission is very important if it has been actively selected. Those workers who have chosen to contribute to a social cause outperform the ones randomly assigned to the same job by about half a standard deviation. This effect seems to be a universal phenomenon that is not driven by information about the alternative job, the choice itself, or a particular subgroup.

Keywords

active decision cognitive dissonance theory field experiment performance prosocial work

Introduction

A wide range of empirical literature has found many differences in organizational characteristics, jobs, and workers between the private and the nonprofit sector (see, e.g., De Cooman et al., 2011; Winter & Thaler, 2016). One difference is that employees of nonprofit organizations (NPOs) are said to care about the social mission of their job (Weisbrod, 1998). Handy and Katz (1998) suggest that lower wages (see, e.g., Borzaga & Tortia, 2006) are used by nonprofits as a screening device to attract intrinsically motivated individuals who are willing to forego some money in exchange for the opportunity to provide goods or services with positive social externalities.

Goal theory of motivation (Locke & Latham, 1990) suggests that employees are more motivated to perform a task if they feel that it will result in important outcomes. Starting from this point, this article studies the conditions under which a prosocial mission can incentivize an existing workforce. With observational data, it is difficult to isolate potential influencing factors on performance due to worker self-selection and measurement difficulties, so we used a natural field experiment (Harrison & List, 2004). In partnership with a local advertising agency, we hired 246 temporary workers, paid them a fixed wage, and measured their performance in doing the simple, homogeneous task of placing solicitation letters—either for businesses or for prosocial organizations—into envelopes and getting them ready to mail. The jobs differed only in the purpose of the letters (prosocial cause vs. for-profit goal) and the allocation method of the jobs (self-selection vs. exogenous allocation).

The broad question of whether a prosocial mission has a positive performance effect can be divided into two narrower questions. First, there is the question of whether a prosocial mission in itself has a positive effect. We investigate it by randomly sorting workers into a mission-oriented and a commercial job. Second, there is the question of whether actively choosing a mission-oriented job has a performance-enhancing effect. Hence, we let a group of workers consciously make the choice to work toward a social mission, and we compared them with workers whom we randomly assigned to the prosocial mission. By offering equal pay, we ensured that almost all workers would prefer the mission over the commercial purpose, so that the sorting effects would be negligible.

Our data suggest that the scope for performance increases through the exogenous provision of a mission is limited. If it exists at all, the pure mission effect is only observable for workers who privately engage in volunteering activities; we gathered this information as a proxy for individuals’ prosociality. We do find, however, a positive and highly significant performance effect caused by the conscious decision to work for a good cause. This performance increase equals about half a standard deviation, and unlike before, it is not driven by a particular subgroup. Further control treatments rule out the possibility that information about the alternative job or the choice itself caused the effect.

We consider our investigation as relevant, since hitherto there is a lack of evidence on the performance effects of prosocial missions from controlled settings. By conducting a field experiment in partnership with profit and NPOs, we are able to rigorously test this fundamental relationship. Moreover, we contribute to the general understanding of the interplay between mission and performance by studying the specific impact of deliberation about a prosocial mission, which we show is more powerful than deliberation and choice in general.

Related Literature and Hypotheses

The suggestion that individuals perform better when working for a social cause than when purely motivated by profit is based on the psychological framework of the goal theory of motivation (Locke & Latham, 1990), which posits that efforts increase with performance goals if employees believe that these will result in important outcomes. Wright (2007) applies this framework to public service motivation and finds that they are more motivated to perform their task if they feel that it is important and achievable. Similarly, the organizational mission influences work motivation by increasing the importance employees place on their work, which can even facilitate innovation (McDonald, 2007). Wright’s analysis is restricted to public employees; applying this theory to our research question, one might argue that a social cause could be perceived as more important than increasing sales and, hence, cause higher performance.

The empirical literature on the relationship between an organization’s mission and workers’ performance can be divided into two strands. The first strand highlights the importance of a mission match. Carpenter and Gong (2016) conducted a real-effort laboratory experiment in which subjects were randomly allocated to two organizations whose missions were diametrically opposed. The performance of mission-matched subjects was more than 70% higher than the performance of mismatched individuals. A similar design was implemented by Smith (2016), who argues that meaningfulness serves as an intervening mechanism; this argument is connected to a recent strand of literature that analyzes the impact of perceived meaning of work on performance (see, e.g., Chadi et al., 2017; Kosfeld et al., 2017). Other experimental studies refrain from explicitly creating matches and mismatches, but they vary the match quality and compare subjects’ performance to a control group in which no mission is present at all. Except for Cassar (2019), most of these studies confirm that there is a positive relationship between mission match and exerted or stated effort (e.g., Gerhards, 2015; Koppel & Regner, 2014; Resh et al., 2017).

The second strand of empirical literature focuses on the impact of social incentives on individuals’ performance, independent of how much the mission is appreciated. By comparing groups of subjects whose efforts either generate donations to a charity or do not, a significant difference in performance can be observed (Kajackaite & Sliwka, 2017), but these effects are often confined to particular settings such as using low stakes (Charness et al., 2016) or subgroups such as low productivity individuals or females (Tonin & Vlassopoulos, 2010, 2015). The experimental designs of these studies are well suited for answering the question of whether additionally pursuing a prosocial mission—for example, in terms of corporate social responsibility activities—enhances workers’ motivation. However, the findings are only incompletely transferable to our question of whether nonprofit workers exert higher efforts than for-profit workers. Note that pursuing a commercial goal can be perceived to be meaningful, too. Then, if the mission effect is mediated by the perceived meaningfulness of work as Smith (2016) suggests, no mission effect might be found. The following empirical and experimental evidence supports this suggestion.

First, findings from the British Household Panel Survey indicate that the likelihood of doing unpaid overtime remains stable if individuals switch from the nonprofit to the for-profit sector (Gregg et al., 2011). Second, the experimental design by Fehrler and Kosfeld (2014) is close to that of the investigations mentioned above, but their control group differs in that subjects’ chosen effort also generates money for a randomly selected student from the university. Compared to this control group, the authors do not find a positive effect of generating donations to a charity, even though subjects were allowed to choose the charity to which donations shall be made. Finally, Banuri et al. (2018) experimentally demonstrate a difference between task-based and mission-based motivation. They show that mission-based motivation enhances performance only if there is no task motivation.

By randomly allocating workers to different solicitation letters—prosocial mission versus commercial advertisements—, we can cleanly identify a pure mission effect on workers’ exerted efforts if it exists. Based on the empirical evidence, we derive the following hypotheses:

Hypothesis 1a: There is no pure mission effect for the overall sample.

Hypothesis 1b: There is a pure mission effect for individuals who regularly volunteer.

Hypothesis 1b suggests that volunteering individuals may be a subgroup of people who exert higher efforts if they are involved with a prosocial mission since prosociality has been shown to be an important factor in mission-oriented settings (see, e.g., Non et al., 2019; Smith, 2016). Second, Weiss and Piderit (1999) find that employees perceive organizational goals to be personally meaningful if the organization’s mission is congruent with their values. One might assume that volunteers are especially likely to value the work of the sender of the mission-oriented letter, an agency that organizes volunteers to do social work in the community. Drawing again on goal theory, the perceived meaningfulness of the organizational goal is likely to translate into observable performance effects.

In reality, jobs are usually not randomly assigned. Even if people do not have a real choice—that is, when the job market is very tight—, accepting a job always involves some commitment. By drawing on cognitive dissonance theory, choosing a mission-oriented job may lead to behavioral changes. As Festinger (1957, p. 2) argues, even though people strive for internal psychological consistency by acting in alignment with their values, “for one reason or another, attempts to achieve consistency may fail.” As a consequence, a feeling of discomfort can arise. However, dissonance caused by a decision—such as taking a job supporting a social cause when one is not committed to that cause—can be reduced by persuading oneself that the chosen alternative is more attractive and/or the rejected alternative is less attractive. Izuma et al. (2010) provide evidence that individuals indeed change their preferences to match their prior decision, using both self-reported preferences and their neural representations. In our setting, the dissonance-reducing activity is the act of putting extra effort into the chosen job because it is not unpleasant or because the social goal is even more desirable than originally thought (see Harmon-Jones et al., 2009). This suggestion is in line with the finding that self-persuasion interventions have a positive effect on job performance; Bellé (2013) reached that conclusion at the end of a study in which nurses had to describe how a specific project could improve the lives of their patients.

Investigating the possible interaction effect between a social mission and the choice of it also follows a recent strand of literature that stresses the importance of self-reflection to behavioral changes. Actively deciding between different jobs might induce workers to reflect on the characteristics of these jobs. In doing so, workers could form a subjective value for the prosocial job (Stutzer et al., 2011), above and beyond a pure mission effect. Relatedly, Krupka and Weber (2009) provide experimental data that individuals’ prosocial behavior increases when their attention is drawn to social norms.

By allowing workers to deliberately sort themselves into a mission-oriented job while we keep the composition of our workforce nearly unchanged due to equal pay, we can investigate the possible existence of what we can call the “active mission choice effect”—a potentially important factor that has been neglected in the mission-based motivation literature so far. Based on the theories and findings we have discussed above, we derive the following hypotheses:

Hypothesis 2a: There is an active mission choice effect for the overall sample.

Hypothesis 2b: The active mission choice effect is greater than the pure mission effect for non-volunteering individuals only.

The latter hypothesis should be true if the active mission choice effect can be explained by the cognitive dissonance theory. In that case, choosing the mission-oriented job does not create any cognitive dissonance for prosocial individuals, so the active choice should not play any role for these individuals. The less socially minded people, however, are in a state of cognitive dissonance and can be expected to put extra effort into the chosen job to avoid discomfort. As a consequence, the effect caused by deliberation in a setting that involves a prosocial mission is likely to be different from the effect caused by the task choice itself. Given that the positive performance effects of task choice usually work via goal commitment (which is a function of the probability that individuals can demonstrate high ability; see, e.g., Nicholls, 1984), choosing between two different commercial advertisements while conducting exactly the same activity should not be related to proving high ability and, hence, should not result in a statistically significant difference in performance.

Study Design

We partnered with a local advertising agency to investigate the impact of a prosocial mission and the conscious choice to contribute to a social cause. Various upcoming mailing campaigns for different nonprofit and for-profit organizations allowed us to conduct a natural field experiment (Harrison & List, 2004) to observe temporary staff in a controlled but natural working environment without the workers being aware of the experimental conduct. King et al. (2013) argue that organizational scholars can derive meaningful conclusions from this methodological approach if the attributes that characterize high-quality field experiments (i.e., random assignment of subjects, a natural setting, and the manipulation of behavioral variables) are present. Participants in natural settings outside the constraints of the laboratory or a survey who are unaware that they are involved in a study can be expected to perform actions that are less filtered by social desirability―which might be an issue especially for nonprofit research. Although field experiments have rarely been the method of choice for scholars publishing in nonprofit journals (Mason, 2013),¹ they have yielded important conclusions for this area, such as about the impact of social information on volunteering (Moseley et al., 2018).

For this study, we attracted the attention of prospective workers using small advertisements on bulletin boards in supermarkets, public libraries, university campuses, and the like, as well as on regional online platforms. Advertisements informed potential workers about the type of job (clerical), the pay, and the contact details. We noted that it was a joint project between our university and local companies and organizations but gave no information regarding the project’s purpose. Interested individuals applied in person during office hours. Some of them brought résumés, while others filled in a short form to provide their contact details, nationality, and age. We briefly informed them about the mailing campaigns, asked them about potential working times, and, if they had brought no résumé, we asked them a few questions about their experience in similar jobs, as is usual in job interviews. Given that we intended to use volunteering as a proxy for prosocial preferences, we asked whether they had engaged in volunteering activities, and, if so, for how long. We used a binary coding system, counting every individual who had volunteered for at least 6 months as a volunteer.²

Afterward, we offered all applicants an immediate opportunity to earn 5 euros for 30 minutes of work. The task consisted of putting letters in envelopes, stamping the envelopes on the front and back, and binding the letters together in stacks of 10 with a rubber band. As the task was the same for all applicants, and as these letters differed from the subsequent treatment letters only in their neutral content—they promoted services for fans of a local sports club—, we used the number of processed letters as a performance indicator.

Of the 267 people who applied for the job and processed the sports information letters, 246 showed up for the allocated work shift. For the 2-hour shift, each received a fixed payment of 20 euros. Upon arrival, a research assistant welcomed the workers and gave them brief standardized instructions for the task at hand. To rule out peer effects, each worker was allocated a different meeting time so that the instructor briefed only one worker at a time. Furthermore, employees worked alone in single offices without any coworkers or supervisors around. All offices were identically equipped with a desk, two office chairs, and about 400 letters and envelopes.³ Workers were also told that they could take breaks whenever necessary. Given these particular circumstances, workers were likely to feel responsible for managing their own workload. In addition, workers were told in advance that this job was a unique opportunity to earn money with this employer to rule out any career concerns.

After the application and before running the experiment, workers were randomly allocated into one out of five treatment groups: Profit, Mission, Choice, Choice Control, and Information Control (for an overview, see Table 1). The task was always identical but the content of the letter and the allocation method differed between treatments. In treatment Profit, the letter served a commercial purpose (advertisements for local shops), whereas, in treatment Mission, the letter requested volunteers for local charities. In both treatments, the content of the letter was given exogenously, so workers had no choice. In treatment Choice, we asked workers which type of letter they preferred to handle. Depending on their decision, we name the groups Choice(Mission) or Choice(Profit), respectively.⁴

Table 1.

Treatment Groups Overview.

1A. Main treatment conditions.
Allocation methodLetter content	Letter content given exogenously	Letter content chosen endogenously by workers
For-profit goal (increasing sales)	Profit	Choice(Profit)
Social cause (recruiting volunteers for a charity)	Mission	Choice(Mission)
1B. Additional control treatments.
Choice Control	Choice between two for-profit letters: Choice Control(Letter 1) and Choice Control(Letter 2).
Information Control	Exogenous allocation of prosocial letters with information that previous workers had enveloped for-profit letters

To investigate whether there is a pure mission effect, we compare the groups Profit and Mission. By comparing Mission and Choice(Mission), we can investigate whether the active and conscious choice of such a mission further enhances workers’ performance. Since lower payments in mission-oriented jobs have been shown to result in strong sorting effects (Fehrler & Kosfeld, 2014), equal pay is a crucial element in our experimental design: As long as working for a prosocial mission is costless, we expected the vast majority of workers to choose this option, so that the characteristics of the workforce in Choice(Mission) should barely differ from the randomly allocated workers in Mission. Theoretically, the comparison between the two groups of Choice(Mission) and Choice(Profit) makes it possible for us to disentangle (a) a performance change due to the respondents’ consciously choosing to do good from (b) a performance change due to the respondents’ simply being given a choice. Given the extremely low sample size in Choice(Profit), we implemented an additional control group, Choice control, in which workers could decide between two different for-profit letters. This control treatment resembles the choice treatment, in that workers were also allowed to make a decision between two options regarding the content of the letters to be processed, but only in Choice could workers deliberately opt for prosocial work.

Since the decision about which option to choose is connected to the information about the different choices, any possible differences in performance might be driven by both the decision and the information. Therefore, we implemented a further control group, Information Control, which differed from treatment Mission only in the factor of information: these workers processed charity letters but were informed that others had already processed for-profit letters.

We thoroughly followed the standard rules for natural field experiments (Harrison & List, 2004), including the strict requirement that we not deceive the workers: their efforts were commercially useful, in terms of processed letters that workers were paid for. Hence, unlike experiments in psychology where deception is the rule rather than the exception, we did not provide a debriefing. Furthermore, the temporary workers were treated according to the rules of German labor law and the intervention was very mild, so that no harm was to be expected; no worker reported any inconvenience. We conducted the experiment in a medium-sized German city in late 2014. All five treatments were rolled out simultaneously, which was possible because of our mild intervention and the naturalness of the situation since different workers might also work on different projects.

Empirical Analysis

The average age of the 246 workers was 24.70 years. 69.5% were female and 11.4% were foreign nationals. About 34.6% had volunteered in the past for at least 6 months, which is consistent with the data from the German Survey on Volunteering (Hagen & Vogel, 2012). We find no statistically significant differences in workers’ initial performance, their age, and volunteering activities between treatments (based on randomization inference with 1,000 random permutations). However, treatment Choice Control includes significantly more female and fewer foreign workers than do the treatments Profit and Choice (see Appendix Table A1). We controlled for this difference using regression analyses.

Looking at workers’ performance, we find that workers processing letters for a for-profit purpose (i.e., treatment Profit) handled an average of 168.53 letters during their shift, while workers exogenously endowed with a mission (treatment Mission) performed equally well, with an average of 172.83 letters (p=.637).⁵ Since individual ability does not differ significantly between treatments (p = .866), we can conclude that there is no pure motivational effect of the mission itself. This finding confirms Hypothesis 1a. To test Hypothesis 1b, we first look at individuals with less than 6 months of volunteering activity. Again, we find no difference between Mission and Profit (p = .385). In contrast, volunteers—those with 6 months or more of volunteer activity—do seem to be positively affected by the prosocial mission. Volunteers working with the letter aimed at recruiting volunteers processed an average of 204.64 letters, while those with the for-profit letter, the average number was only 172.64. This results in an economically significant raw difference of almost 18% (p = .093, but with only 14 observations each). Regression analyses (see Table 2), which control for potential differences in initial productivity between volunteers and nonvolunteers, support these findings.

Result 1: We do not observe a pure mission effect when comparing a mission-oriented and a for-profit job. If an economically significant impact exists at all, it can be reported only for regularly engaged volunteers. These findings support both Hypothesis 1a and Hypothesis 1b.

We now turn to the potential “active mission choice effect.” In the treatment Choice, in which workers could choose the type of letter they wanted to process, whether mission-oriented or for-profit, without any impact on their pay, the vast majority (87%) opted, as we expected, for the prosocial mission. The average performance in Choice(Mission), 195.75 letters, is significantly higher than in both previous treatments in which the letters were assigned to the workers (Profit: p = .010; Mission: p = .052). Compared to the workers which were assigned to the commercial letter, this is a raw difference of about 16.2%, which equals more than half a standard deviation (Hedges’s g_s = 0.575, 95% CI = [0.098, 1.053]).

Table 2.

Regression Analysis for the Pure Mission Effect.

	(1)	(2)	(3)	(4)
Profit	Ref.	Ref.	Ref.	Ref.
Mission	4.300 (11.057)	3.518 (6.866)	3.359 (7.029)	−6.081 (8.103)
Initial performance		3.126*** (0.258)	3.030*** (0.282)	2.964*** (0.269)
Female			4.208 (6.977)	5.766 (7.077)
Age			−0.054 (0.404)	−0.106 (0.378)
Foreign			−1.241 (9.347)	−1.753 (9.307)
Volunteer			8.870 (7.897)	−4.402 (10.961)
Mission × volunteer				26.694* (15.628)
Constant	168.525*** (7.393)	51.916*** (9.530)	51.273*** (17.268)	58.786*** (16.008)
Observations	80	80	80	80
Adjusted R²	—	.611	.600	.613

Note. Dependent variable is the total number of enveloped letters during the 2-hour shift. The table reports OLS coefficient estimates (robust standard errors in parentheses). Only workers from the treatments Mission and Profit are considered. OLS = ordinary least squares.

p < .1. ***p < .01.

Since not all of the workers allocated to the treatment Choice opted for the prosocial cause, this difference cannot be interpreted as the average treatment effect (ATE) we usually would like to estimate. Hence, in the first step, we want to exclude the possibility that the performance increase is not a real treatment effect but is only driven by a change in the workforce’s composition, such as especially high ability individuals sorting into the mission-oriented job. Figure 1 illustrates that both the means and the distribution of initial performance are comparable across treatments, whereas the distribution of the workers’ output in the group Choice(Mission) is clearly shifted to the right after treatment intervention. This suggests that the treatment effect is not driven by single outliers and, hence, is unlikely to be caused by a change in the workforce’s composition. The exclusion of one worker who performed substantially better than all remaining ones during the first 30-minute shift does qualitatively not change any of our results.

Figure 1.

Cumulative distribution functions of workers’ performance.

Given the somewhat higher share of volunteers compared to the groups Profit and Mission (42.5% vs. 35%) and that the pure mission effect can be found in prosocial individuals only (if it exists at all), we use regression analysis to include an interaction term for having actively chosen the prosocial mission and being a volunteer (see Table 3). Again, the results contradict the notion that the observed treatment effect is driven by a change in the workforce’s composition because the point estimate for Choice(Mission) remains nearly unchanged and the included interaction term is positive but insignificant. We also checked for the heterogeneity of the treatment effect as regards individual ability or workers’ gender, but the active mission choice effect does not seem to be restricted to any specific subgroup.

Table 3.

The Active Mission Choice Effect—Testing for Heterogeneity.

	(1)	(2)	(3)	(4)
Profit	Ref.	Ref.	Ref.	Ref.
Choice (Mission)	26.121 (7.081)	23.135*** (7.934)	23.137** (11.417)	28.461*** (8.849)
Volunteer	1.258 (7.622)	−2.886 (11.139)	1.568 (7.764)	1.133 (7.745)
Choice (Mission) × volunteer		7.767 (15.837)
Female	−0.026 (7.380)	0.531 (7.500)	−2.521 (9.112)	0.492 (7.751)
Choice (Mission) × female			4.573 (14.142)
High ability workers				6.622 (15.169)
Choice (Mission) × high ability				−4.572 (15.061)
Constant	60.905*** (18.602)	61.315*** (18.723)	63.004*** (18.612)	62.922*** (21.122)
Observations	80	80	80	80
Adjusted R²	.599	.595	.594	.589

Note. Dependent variable is the total number of enveloped letters during the 2-hour shift. The table reports OLS coefficient estimates (robust standard in parentheses). Only workers from the treatments Profit and Choice (Mission) are considered. Additional controls include workers’ age, nationality, and initial performance. OLS = ordinary least squares.

**p < .05. ***p < .01.

Taken together, our data suggest that the observed effect is not due to self-selection of workers who would perform well anyway but that workers exert higher efforts if they deliberately decided upon doing something good, which supports Hypothesis 2a. Since deciding in favor of the commercial advertisement can be interpreted as noncompliance from a statistical point of view, we finally apply a two-stage least squares estimation (2SLS) to estimate the complier average causal effect (CACE), as proposed by Gerber and Green (2012), which is the consistent estimate of the ATE among the particular subset of individuals who have chosen the prosocial job. In the first-stage regression, none of the controls is significant, and the treatment is estimated to raise workers’ output by 25.62 letters (p = .000) compared to the exogenous allocation of the for-profit letter. This is only minimally smaller than the raw difference of 27.22 letters, confirming our previous suggestion that the treatment effect equals about half a standard deviation. The results remain nearly identical when using treatment Mission as the reference category.

To scrutinize the suggestion that a feasible channel of the active mission choice effect is a result of individuals’ avoidance of cognitive dissonances, we again split the sample according to workers’ volunteering activities. We find no performance difference between volunteers and nonvolunteers in the group Choice(Mission) itself (p = .956), nor do we observe a difference between volunteers who were exogenously allocated the mission-oriented job and those who actively decided in favor of working for the prosocial cause (p = .416). However, we find a highly significant treatment effect of actively choosing to do good for the subsample of nonvolunteers (155.69 vs. 197.48 letters, p = .004). This observation suggests that the less socially minded individuals indeed seem to increase their effort to avoid cognitive dissonance.

Result 2: In contrast to an exogenously provided mission, the active choice to do good sizably increases the work performance of the whole group. Given that only nonvolunteers are sensitive to this specific treatment, our data suggest that the avoidance of cognitive dissonances is a likely behavioral driver.

To rule out the possibility that the previously identified effect is driven by information about the different options or the mere right to choose—which are two inevitable aspects of self-selection—we analyze the work performance in the control treatments, Information Control and Choice Control. Given that the workers who opted for the for-profit instead of the charity letters in treatment Choice perform on average (166.33 letters) almost identically to workers who were exogenously assigned to one of the letters, it is rather unlikely that the choice itself boosts employees’ performance, even though this number is based on only six workers. Both control groups, Choice Control(Letter 1) and Choice Control(Letter 2), with an average of 167.87 and 168.83 enveloped letters respectively, confirm this suggestion.⁶

Finally, being aware of the existence of the for-profit letters in the treatment Information Control yields the highest average of 176.30 enveloped letters among all experimental groups besides Choice (Mission). This number, however, is neither significantly different from treatment Profit (p = .196) nor Mission (p = .519). Both the regression analyses (see Table 4) and the multiple hypothesis testing, as proposed by List et al. (2019), resulting in p = .043, confirm that only the treatment Choice(Mission) clearly outperformed the treatment Profit.

Table 4.

Treatment Effects Overview.

	(1)	(2)
Profit	Ref.	Ref.
Mission	4.300 (11.076)	3.363 (6.875)
Choice (Mission)	27.225** (10.500)	25.833*** (6.846)
Choice (Profit)	−2.192 (17.245)	−2.029 (9.299)
Information Control	7.775 (9.265)	6.616 (6.722)
Choice Control (Letter 1)	−0.657 (9.225)	−3.855 (5.589)
Choice Control (Letter 2)	0.308 (10.050)	−1.226 (5.995)
Constant	168.525*** (7.406)	56.264*** (11.211)
Additional controls	No	Yes
Observations	246	244
Adjusted R²	.024	.595

Note. Dependent variable is the total number of enveloped letters during the 2-hour shift. Additional controls include workers’ age, gender, nationality, initial performance, and whether they have been volunteering before. In Specification (2), two observations are lost due to missings as regards their birthdate. The table reports OLS coefficient estimates (robust standard in parentheses). Comparing the point estimates for Mission(Choice) and Information Control, which are closest to each other, we still find a highly significant difference between the two (p = .007). OLS = ordinary least squares.

p < .05. ***p < .01.

Discussion

In an ideal world, employees have a clear idea of their preferred mission and the organizations they would like to work for. They search until they find the perfect job, with prosocial individuals reliably sorting themselves into mission-oriented jobs and NPOs. Actually, it is not uncommon that people have less precise ideas about their preferences. To put it positively, they are open to new experiences and can imagine working for either a prosocial or a profit-oriented organization. Sometimes, they need to find a job quickly, and they take the first-best they can get, so that chance determines where they go.

The existing literature cannot adequately answer the question of whether a job’s prosocial mission per se has a positive impact on workers’ willingness to exert effort compared to a profit goal. Our study seeks to fill this gap by providing clean, causal evidence obtained from a natural field experiment. We find that employees allocated to the prosocial mission work as hard as those working toward a profit goal. Only when looking at a very specific subgroup, namely people who already do unpaid volunteer work during their leisure time, is a pure mission effect observable. This coincides with the assumption that a prosocial mission is more important for these people than for the average population. Note, however, that we aimed to isolate the mission effect from any other influencing factors. Consequently, the mission of the job was communicated in a very neutral way, without highlighting its value for society, but in ordinary work situations, organizations could draw on active leadership communication to inspire individuals to work harder for their social cause (Jensen et al., 2018). Especially where the working environments are monotonous, Antonakis et al. (2019) have shown that charismatic leadership has a significant impact on workers’ performance.

Besides exogenously allocating them into treatment groups, we allowed some employees in different treatment conditions to opt for their preferred job because self-selection is an inherent part of the real labor market. The innovation here is that workers could choose between a normal commercial job and doing good for society without an earnings loss. Thereby, we were able to generate a particularly interesting group for research: a workforce that deliberately chose a job with a prosocial mission without the usually biased workforce characteristics due to selection. We find that there is no additional performance effect of the active choice for the subgroup of volunteer workers who already performed very well under the exogenous allocation of the prosocial mission. We do, however, observe a statistically significant performance increase for nonvolunteers who actively decide in favor of the social cause, which suggests that the avoidance of cognitive dissonance is a likely behavioral driver.

While previous research emphasizes the importance of a perfect mission match, and therefore suggests that NPO managers should pay particular attention to perfect screening and selection, our results suggest an overestimation of a perfect mission match, given that we observe a positive performance effect once individuals decide to work for such a social cause regardless of their preferences. The occurrence of this active mission choice effect seems to be good news for society as a whole but also for human resource management. It confirms that human resource managers in NPOs are indeed well advised to pursue the strategy of having a values-laden orientation, as shown by Ridder and McCandless (2010). It is, however, important that NPOs are aware of this resource and communicate the mission accordingly also outside the organization so that potential workers can actively decide to contribute to a good cause.

This finding is interesting and novel, but it should be followed by further investigations involving larger numbers of workers. While the statistical power of our investigations is completely sufficient regarding this “active mission choice effect” (0.738), the subgroup analyses that specifically study the behavior of workers classified as prosocial would benefit from larger samples in terms of confidence in these novel scientific contributions.

Even though we intentionally designed our experiment in such a way that we can rule out any influences from external rewards, a future study could analyze to what extent the perceived fairness of a fixed wage influences the established findings. Given the present setting of a one-time job without any specific requirements applicants had to fulfill, the hourly wage can be perceived as rather generous. In contrast, agents employed at nonprofit organizations usually receive significantly lower salaries than do employees working for profit-oriented companies. If the positive performance effect of actively choosing to do good was tied to perceiving the paid wage as fair, it would be interesting for employers to think about wage increases, at least moderate ones, ending up in a beneficial situation for both employee and employer. The evidence provided by Brown and Yoshioka (2003), showing that dissatisfaction with pay tended to override employees’ attitudes toward the mission which is otherwise positively associated with their intentions to remain within the organization, emphasizes the need to investigate the role of pay.

Given that the active mission choice effect seems to be driven by people who work harder to strive for internal psychological consistency, future research should also aim at investigating the long-run consequences of such behavior. On one hand, individuals’ desire to avoid cognitive dissonance might fatigue them so that they become more likely to reduce their effort or even quit the job. On the other hand, however, such avoidance behavior might not be necessary in the long run for at least two reasons. First, Stutzer et al. (2011) indicate that attributing value to a particular prosocial activity may require introspection, but many people seem to refrain from doing so. If they start working for a prosocial cause, organizational socialization might make it necessary to introspect and altruistic preferences could adapt. Second, some individuals might not consider certain social activities as particularly important because they might not have the necessary information to assess their importance. They could even have prejudices toward potential clients or coworkers. According to Allport’s contact hypothesis (also known as intergroup contact theory), contact with an unfamiliar outgroup improves knowledge, increases empathy, and reduces fears, ultimately breaking down prejudices (see, e.g., the meta-analysis by Pettigrew & Tropp, 2006). As a consequence, accepting a mission-oriented job could (positively) change individuals’ perceptions by getting in contact with the matter. Using both lines of reasoning, one might argue that avoiding cognitive dissonances might not be necessary in the long run.

In that we conducted a highly controlled field experiment, causal inference can be made and internal validity is high. Compared to laboratory studies, our research design provides high external validity, too, because we observed real workers while they were performing a genuine task. An experimental approach not only provides particularly reliable and valid evidence but is also very rare in the field of intrinsic motivation research (see Cerasoli et al., 2014)―even though already Grant (2008) proposed that future research should experimentally manipulate prosocial and intrinsic motivations to provide more conclusive evidence on this topic. This ongoing paucity of experimental studies is problematic in that Cerasoli et al. (2014) also show that experimental studies find weaker associations between intrinsic motivation and performance than correlational designs, a result that we can confirm with our data.

We acknowledge that our particular setting, in which short-term workers conduct a single monotonous, routine task, may raise the question of the generalizability of the results. At this point, two counterarguments are warranted. First, we test general theories, which should be valid (or invalid, as the case may be) in this setting, as they are elsewhere. Second, this setting is not just any setting, but one that is often found in reality. We are not examining a convenience sample, but exactly those who have applied for temporary employment. In other words, precisely those employees who are many organizations and companies have to rely on from time to time to reduce high workloads, and who are often difficult to motivate to do their best due to the short duration of the employment and the modest pay.

Still, further research should show whether our results are transferable to full-time employees. Usually, full-time employees have a varying set of tasks with which they might not directly contribute to the organizational mission. Hence, it would be interesting to know whether the performance effects observed here are limited to tasks that are directly related to the prosocial cause; whether such tasks can have a positive spillover on unrelated tasks in the event that both are present; and whether the organizational mission might be enough for employees who are barely in touch with the mission, such as those working in an organization’s administration taking care of its financial performance. Given that Knapp et al. (2017) do not find that task significance has a major impact on nonprofit employees’ job satisfaction and turnover intention, our findings are likely to carry over to other settings, but further research in this context would contribute to a deeper understanding of how a prosocial mission influences individual behavior at work.

Footnotes

Appendix

Table A1.

Descriptive Statistics by Treatment Group.

Random allocation into:	Profit (N = 40)	Mission (N = 40)	Choice (N = 46)	Choice Control (N = 80)	Information Control (N = 40)	Dropouts (N = 21)
Initial performance	37.30	37.55	37.61	38.01	37.70	34.57
Performance during 2-hour shift	168.53	172.83	191.91	168.34	176.3	—
Socio-demographics:
Female	0.650	0.675	0.630	0.788	0.650	0.619
Age	25.70	24.25	24.22	24.79	24.34	25.56
Foreign	0.175	0.175	0.152	0.051	0.075	0.059
Regular volunteers	0.350	0.350	0.391	0.313	0.350	0.381

Self-selection into:	Choice (Mission) (N = 40)	Choice (Profit) (N = 6)	Choice Control (Letter 1) (N = 38)	Choice Control (Letter 2) (N = 42)
Initial performance	37.63	37.5	38.42	37.64
Performance during 2-hour shift	195.75	166.33	167.87	168.83
Socio-demographics:
Female	0.625	0.667	0.684	0.881
Age	24.25	24.00	25.30	24.34
Foreign	0.150	0.167	0.054	0.048
Regular volunteers	0.425	0.167	0.342	0.286

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sabrina Jeworrek

Notes

Author Biographies

Sabrina Jeworrek is assistant professor for applied microeconometrics at Otto von Guericke University Magdeburg and leader of the research group organizational behavior and corporate success at the Halle Institute for Economic Research, Germany. Her research interests deal with the impact of managerial behavior and nonmonetary incentives on work effort, unpaid work, and social preferences.

Vanessa Mertins is professor for the management of social services at the University of Vechta, Germany. Her research covers projects on work motivation for both paid and unpaid work-, gender-, and ethnicity-based discrimination and the promotion of physical activity.

References

Antonakis

d’Adda

Weber

Zehnder

(2019). “Just words? Just speeches?” On the economic value of charismatic leadership [Working Paper]. Université de Lausanne.

Banuri

Keefer

de Walque

(2018). Love the job. . . or the patient? Task vs. mission-based motivations in health care (World Bank Group Policy Research Working Paper 8338). World Bank.

Bellé

(2013). Experimental evidence on the relationship between public service motivation and job performance. Public Administration Review, 731(1), 143–153.

Borzaga

Tortia

(2006). Worker motivations, job satisfaction, and loyalty in public and nonprofit social services. Nonprofit and Voluntary Sector Quarterly, 35(2), 225–248.

Brown

W. A.

Yoshioka

C. F.

(2003). Mission attachment and satisfaction as factors in employee retention. Nonprofit Management and Leadership, 14(1), 5–18.

Carpenter

Gong

(2016). Motivating agents: How much does the mission matter? Journal of Labor Economics, 34(1), 211–236.

Cassar

(2019). Job mission as a substitute for monetary incentives: Benefits and limits. Management Science, 65(2), 896–912.

Cerasoli

C. P.

Nicklin

J. M.

Ford

M. T.

(2014). Intrinsic motivation and extrinsic incentives jointly predict performance: A 40-year meta-analysis. Psychological Bulletin, 140(4), 980–1008.

Chadi

Jeworrek

Mertins

(2017). When the meaning of work has disappeared: Experimental evidence on employees’ performance and emotions. Management Science, 63(6), 1696–1707.

10.

Charness

Cobo-Reyes

Sánchez

(2016). The effect of charitable giving on workers’ performance: Experimental evidence. Journal of Economic Behavior & Organization, (Part A), 131, 61–74.

11.

Clary

E. G.

Snyder

Stukas

A. A.

(1996). Volunteers’ motivations: Findings from a national survey. Nonprofit and Voluntary Sector Quarterly, 25(4), 485–505.

12.

De Cooman

De Gieter

Pepermans

Jegers

(2011). A cross-sector comparison of motivation-related concepts in for-profit and not-for-profit service organizations. Nonprofit and Voluntary Sector Quarterly, 40(2), 296–317.

13.

Fehrler

Kosfeld

(2014). Pro-social missions and worker motivation: An experimental study. Journal of Economic Behavior & Organization, 100, 99–110.

14.

Festinger

(1957). A theory of cognitive dissonance. Stanford University Press.

15.

Gerber

A. S.

Green

D. P.

(2012). Field experiments: Design, analysis, and interpretation. W. W. Norton.

16.

Gerhards

(2015). The incentive effects of missions—evidence from experiments with NGO employees and students. European Economic Review, 79, 252–262.

17.

Grant

A. M.

(2008). Does intrinsic motivation fuel the prosocial fire? Motivational synergy in predicting persistence, performance, and productivity. Journal of Applied Psychology, 93(1), 48–58.

18.

Gregg

Grout

P. A.

Ratcliffe

Smith

Windmeijer

(2011). How important is pro-social behavior in the delivery of public services? Journal of Public Economics, 95(7-8), 758–766.

19.

Hagen

Vogel

(2012). Freiwilliges und generationenübergreifendes Engagement von Frauen und Männern—Analysepotenziale und Weiterentwicklung des Deutschen Freiwilligen-surveys. Informationsdienst Altersfrage, 39(3), 3–9.

20.

Handy

Katz

(1998). The wage differential between nonprofit institutions and corporations: Getting more by paying less? Journal of Comparative Economics, 26(2), 246–261.

21.

Harmon-Jones

Amodio

D. M.

Harmon-Jones

(2009). Action-based model of dissonance: A review, integration, and expansion of conceptions of cognitive conflict. In Zanna

M. P.

(Ed.), Advances in experimental social psychology(Vol. 41, pp. 119–166). Academic Press.

22.

Harrison

G. W.

List

J. A.

(2004). Field experiments. Journal of Economic Literature, 42(4), 1009–1055.

23.

Heinz

Schumacher

(2017). Signaling cooperation. European Economic Review, 98, 199–216.

24.

Izuma

Matsumoto

Murayama

Samejima

Sadato

Matsumoto

(2010). Neural correlates of cognitive dissonance and choice-induced preference change. Proceedings of the National Academy of Sciences, 107(51), 22014–22019.

25.

Jensen

U. T.

Moynihan

D. P.

Houlberg Salomonsen

(2018). Communicating the vision: How face-to-face dialogue facilitates transformational leadership. Public Administration Review, 78(3), 350–361.

26.

Kajackaite

Sliwka

(2017). Social responsibility and incentives in the lab: Why do agents exert more effort when principals donate? Journal of Economic Behavior & Organization, 142, 482–493.

27.

King

E. B.

Hebl

M. R.

Botsford Morgan

Ahmad

A. S.

(2013). Field Experiments on Sensitive Organizational Topics. Organizational Research Methods, 16(4), 501–521.

28.

Knapp

J. R.

Smith

B. R.

Sprinkle

T. A.

(2017). Is it the job or the support? Examining structural and relational predictors of job satisfaction and turnover intention for nonprofit employees. Nonprofit and Voluntary Sector Quarterly, 46(3), 652–671.

29.

Koppel

Regner

(2014). Corporate social responsibility in the workplace. Experimental Economics, 17(3), 347–370.

30.

Kosfeld

Neckermann

Yang

(2017). The effects of financial and recognition incentives across work contexts: The role of meaning. Economic Inquiry, 55(1), 237–247.

31.

Krupka

Weber

R. A.

(2009). The focusing and informational effects of norms on pro-social behavior. The Journal of Economic Psychology, 30(3), 307–320.

32.

List

J. A.

Shaikh

(2019). Multiple hypothesis testing in experimental economics. Experimental Economics, 22, 773–793.

33.

Locke

E. A.

Latham

G. P.

(1990). A theory of goal setting & task performance. Prentice-Hall.

34.

Mason

D. P.

(2013). Putting charity to the test: A case for field experiments on giving time and money in the nonprofit sector. Nonprofit and Voluntary Sector Quarterly, 42(1), 193–202.

35.

McDonald

R. E.

(2007). An investigation of innovation in nonprofit organizations: The role of organizational mission. Nonprofit and Voluntary Sector Quarterly, 36(2), 256–281.

36.

Moseley

James

John

Richardson

Ryan

Stoker

(2018). The effects of social information on volunteering: A field experiment. Nonprofit and Voluntary Sector Quarterly, 47(3), 583–603.

37.

Nicholls

J. G.

(1984). Achievement motivation: Conceptions of ability, subjective car, task choice, and performance. Psychological Review, 913(3), 328–346.

38.

Non

Rohde

de Grip

Dohmen

(2019). Mission of the company, prosocial attitudes and job preferences: A discrete choice experiment (ROA Research Memorandum 2019/6). Maastricht University, Research Centre for Education and the Labour Market.

39.

Pettigrew

T. F.

Tropp

L. R.

(2006). A meta-analytic test of intergroup contact theory. Journal of Personality and Social Psychology, 90(5), 751–789.

40.

Resh

Marvel

Wen

(2017). The persistence of prosocial work effort as a function of mission match. Public Administration Review, 78(1), 116–125.

41.

Ridder

H.-G.

McCandless

(2010). Influences on the architecture of human resource management in nonprofit organization. Nonprofit and Voluntary Sector Quarterly, 39(1), 124–141.

42.

Smith

(2016). The motivational effects of mission matching: A lab-experimental test of a moderated mediation model. Public Administration Review, 76(4), 626–637.

43.

Stutzer

Goette

Zehnder

(2011). Active decisions and prosocial behaviour: A field experiment on blood donation. The Economic Journal, 121(556), 476–493.

44.

Tonin

Vlassopoulos

(2010). Disentangling the sources of pro-socially motivated effort: A field experiment. Journal of Public Economics, 94(11), 1086–1092.

45.

Tonin

Vlassopoulos

(2015). Corporate philanthropy and productivity: Evidence from an online real effort experiment. Management Science, 61(8), 1795–1811.

46.

Weisbrod

B. A.

(1998). The nonprofit economy. Harvard University Press.

47.

Weiss

J. A.

Piderit

S. K.

(1999). The value of mission statements in public agencies. Journal of Public Administration Research and Theory, 9(2), 193–224.

48.

Winter

Thaler

(2016). Does motivation matter for employer choices? A discrete-choice analysis of medical students’ decision among public, nonprofit, and for-profit hospitals. Nonprofit and Voluntary Sector Quarterly, 45(4), 762–786.

49.

Wright

B. E.

(2007). Public service and motivation: Does mission matter? Public Administration Review, 67(1), 54–64.